Research Of PSO Based Pidnn Decoupling Control In Laser Diode

It is particularly important for the temperature and current controlling ofsemiconductor laser diode (LD) system because LD is very sensitive for temperatureand drive current. In the research of LD’s temperature and current control problems,there is certain coupling between temperature and light output power, so this paperfocuses on LD decoupling control for improving the control precision and performanceof LD.In this article, DSP is selected as system control core, and constant currentsource driver circuit, temperature control circuit and isolation protection circuit aredesigned to improve the stability of LD system. Due that thermal electrical cooler(TEC) is the main temperature control executing device, coupling mechanism modelfor LD system is set up from the energy conservation and work principle of TEC.Through studying coupling experiment model under different input signals, themechanism model is verified and revised. Based on LD system model, the paperadopts positive and reverse feed-forward compensation decoupling control methods tosimulating LD control. Considering that commonly overlooked the complexity ofcoupling model and environment instability in building LD mechanism model, weadopt the PIDNN (PID neural network) without the coupling model to simulate LDcontrol. In addition, In order to ensure the convergence and convergence rate ofPID neural network, PSO (particle swarm optimization) learning algorithm is used.Sequentially, the merits and demerits are obtained by comparing curves of temperatureand output power for above decoupling control methods.The experiment results indicate that PID neural network decoupling system withPSO has advantage of higher accuracy, lower overshoot and stronger ability ofself-adaptation to meet the system requirements, which the output optical powerfluctuation of LD’s driving system was successfully kept around0.01W level within25seconds, the temperature fluctuation was kept less than±0.1℃within80secondsunder normal temperature. In the last part of this paper, the existing problems and thefuture ameliorations is discussed.